MLCE-103 Add necessary enhancements to ModelAccuracyTool

* Evaluate model accuracy using category names instead of numerical
labels.
* Add blacklist support
* Add range selection support

Signed-off-by: SiCong Li <sicong.li@arm.com>
Change-Id: I7b1d2d298cfcaa56a27a028147169404b73580bb

2 files changed, 115 insertions, 40 deletions

diff --git a/src/armnnUtils/ModelAccuracyChecker.cpp b/src/armnnUtils/ModelAccuracyChecker.cpp
index bee5ca2365..81942dc2be 100644
--- a/src/armnnUtils/ModelAccuracyChecker.cpp
+++ b/src/armnnUtils/ModelAccuracyChecker.cpp
@@ -3,22 +3,27 @@
 // SPDX-License-Identifier: MIT
 //
 
-#include <vector>
-#include <map>
-#include <boost/log/trivial.hpp>
 #include "ModelAccuracyChecker.hpp"
+#include <boost/filesystem.hpp>
+#include <boost/log/trivial.hpp>
+#include <map>
+#include <vector>
 
 namespace armnnUtils
 {
 
-armnnUtils::ModelAccuracyChecker::ModelAccuracyChecker(const std::map<std::string, int>& validationLabels)
-    : m_GroundTruthLabelSet(validationLabels){}
+armnnUtils::ModelAccuracyChecker::ModelAccuracyChecker(const std::map<std::string, std::string>& validationLabels,
+                                                       const std::vector<LabelCategoryNames>& modelOutputLabels)
+    : m_GroundTruthLabelSet(validationLabels)
+    , m_ModelOutputLabels(modelOutputLabels)
+{}
 
 float ModelAccuracyChecker::GetAccuracy(unsigned int k)
 {
-    if(k > 10) {
-        BOOST_LOG_TRIVIAL(info) << "Accuracy Tool only supports a maximum of Top 10 Accuracy. "
-                                   "Printing Top 10 Accuracy result!";
+    if (k > 10)
+    {
+        BOOST_LOG_TRIVIAL(warning) << "Accuracy Tool only supports a maximum of Top 10 Accuracy. "
+                                      "Printing Top 10 Accuracy result!";
         k = 10;
     }
     unsigned int total = 0;
@@ -28,4 +33,43 @@ float ModelAccuracyChecker::GetAccuracy(unsigned int k)
     }
     return static_cast<float>(total * 100) / static_cast<float>(m_ImagesProcessed);
 }
-}
\ No newline at end of file
+
+// Split a string into tokens by a delimiter
+std::vector<std::string>
+    SplitBy(const std::string& originalString, const std::string& delimiter, bool includeEmptyToken)
+{
+    std::vector<std::string> tokens;
+    size_t cur  = 0;
+    size_t next = 0;
+    while ((next = originalString.find(delimiter, cur)) != std::string::npos)
+    {
+        // Skip empty tokens, unless explicitly stated to include them.
+        if (next - cur > 0 || includeEmptyToken)
+        {
+            tokens.push_back(originalString.substr(cur, next - cur));
+        }
+        cur = next + delimiter.size();
+    }
+    // Get the remaining token
+    // Skip empty tokens, unless explicitly stated to include them.
+    if (originalString.size() - cur > 0 || includeEmptyToken)
+    {
+        tokens.push_back(originalString.substr(cur, originalString.size() - cur));
+    }
+    return tokens;
+}
+
+// Remove any preceding and trailing character specified in the characterSet.
+std::string Strip(const std::string& originalString, const std::string& characterSet)
+{
+    BOOST_ASSERT(!characterSet.empty());
+    const std::size_t firstFound = originalString.find_first_not_of(characterSet);
+    const std::size_t lastFound  = originalString.find_last_not_of(characterSet);
+    // Return empty if the originalString is empty or the originalString contains only to-be-striped characters
+    if (firstFound == std::string::npos || lastFound == std::string::npos)
+    {
+        return "";
+    }
+    return originalString.substr(firstFound, lastFound + 1 - firstFound);
+}
+}    // namespace armnnUtils
\ No newline at end of file
diff --git a/src/armnnUtils/ModelAccuracyChecker.hpp b/src/armnnUtils/ModelAccuracyChecker.hpp
index cdd2af0ac5..c4dd4f1b05 100644
--- a/src/armnnUtils/ModelAccuracyChecker.hpp
+++ b/src/armnnUtils/ModelAccuracyChecker.hpp
@@ -5,39 +5,81 @@
 
 #pragma once
 
+#include <algorithm>
+#include <armnn/Types.hpp>
+#include <boost/assert.hpp>
+#include <boost/variant/apply_visitor.hpp>
 #include <cstddef>
-#include <string>
+#include <functional>
+#include <iostream>
 #include <map>
+#include <string>
 #include <vector>
-#include <boost/variant/apply_visitor.hpp>
-#include <iostream>
-#include <armnn/Types.hpp>
-#include <functional>
-#include <algorithm>
 
 namespace armnnUtils
 {
 
 using namespace armnn;
 
+// Category names associated with a label
+using LabelCategoryNames = std::vector<std::string>;
+
+/** Split a string into tokens by a delimiter
+ *
+ * @param[in] originalString    Original string to be split
+ * @param[in] delimiter         Delimiter used to split \p originalString
+ * @param[in] includeEmptyToekn If true, include empty tokens in the result
+ * @return A vector of tokens split from \p originalString by \delimiter
+ */
+std::vector<std::string>
+    SplitBy(const std::string& originalString, const std::string& delimiter = " ", bool includeEmptyToken = false);
+
+/** Remove any preceding and trailing character specified in the characterSet.
+ *
+ * @param[in] originalString    Original string to be stripped
+ * @param[in] characterSet      Set of characters to be stripped from \p originalString
+ * @return A string stripped of all characters specified in \p characterSet from \p originalString
+ */
+std::string Strip(const std::string& originalString, const std::string& characterSet = " ");
+
 class ModelAccuracyChecker
 {
 public:
-    ModelAccuracyChecker(const std::map<std::string, int>& validationLabelSet);
-
+    /** Constructor for a model top k accuracy checker
+     *
+     * @param[in] validationLabelSet Mapping from names of images to be validated, to category names of their
+                                     corresponding ground-truth labels.
+     * @param[in] modelOutputLabels  Mapping from output nodes to the category names of their corresponding labels
+                                     Note that an output node can have multiple category names.
+     */
+    ModelAccuracyChecker(const std::map<std::string, std::string>& validationLabelSet,
+                         const std::vector<LabelCategoryNames>& modelOutputLabels);
+
+    /** Get Top K accuracy
+     *
+     * @param[in] k The number of top predictions to use for validating the ground-truth label. For example, if \p k is
+                    3, then a prediction is considered correct as long as the ground-truth appears in the top 3
+                    predictions.
+     * @return  The accuracy, according to the top \p k th predictions.
+     */
     float GetAccuracy(unsigned int k);
 
-    template<typename TContainer>
+    /** Record the prediction result of an image
+     *
+     * @param[in] imageName     Name of the image.
+     * @param[in] outputTensor  Output tensor of the network running \p imageName.
+     */
+    template <typename TContainer>
     void AddImageResult(const std::string& imageName, std::vector<TContainer> outputTensor)
     {
         // Increment the total number of images processed
         ++m_ImagesProcessed;
 
         std::map<int, float> confidenceMap;
-        auto & output = outputTensor[0];
+        auto& output = outputTensor[0];
 
         // Create a map of all predictions
-        boost::apply_visitor([&](auto && value)
+        boost::apply_visitor([&confidenceMap](auto && value)
                              {
                                  int index = 0;
                                  for (const auto & o : value)
@@ -64,8 +106,7 @@ public:
         std::set<std::pair<int, float>, Comparator> setOfPredictions(
             confidenceMap.begin(), confidenceMap.end(), compFunctor);
 
-        std::string trimmedName = GetTrimmedImageName(imageName);
-        int value = m_GroundTruthLabelSet.find(trimmedName)->second;
+        const std::string correctLabel = m_GroundTruthLabelSet.at(imageName);
 
         unsigned int index = 1;
         for (std::pair<int, float> element : setOfPredictions)
@@ -74,7 +115,10 @@ public:
             {
                 break;
             }
-            if (element.first == value)
+            // Check if the ground truth label value is included in the topi prediction.
+            // Note that a prediction can have multiple prediction labels.
+            const LabelCategoryNames predictionLabels = m_ModelOutputLabels[static_cast<size_t>(element.first)];
+            if (std::find(predictionLabels.begin(), predictionLabels.end(), correctLabel) != predictionLabels.end())
             {
                 ++m_TopK[index];
                 break;
@@ -83,24 +127,11 @@ public:
         }
     }
 
-    std::string GetTrimmedImageName(const std::string& imageName) const
-    {
-        std::string trimmedName;
-        size_t lastindex = imageName.find_last_of(".");
-        if(lastindex != std::string::npos)
-        {
-            trimmedName = imageName.substr(0, lastindex);
-        } else
-        {
-            trimmedName = imageName;
-        }
-        return trimmedName;
-    }
-
 private:
-    const std::map<std::string, int> m_GroundTruthLabelSet;
-    std::vector<unsigned int> m_TopK = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
-    unsigned int m_ImagesProcessed = 0;
+    const std::map<std::string, std::string> m_GroundTruthLabelSet;
+    const std::vector<LabelCategoryNames> m_ModelOutputLabels;
+    std::vector<unsigned int> m_TopK = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
+    unsigned int m_ImagesProcessed   = 0;
 };
 } //namespace armnnUtils